Backup of only new or changed user data sets is known as incremental backup. Automatic incremental backup ensures that up to date copies of new and changed data sets exist in case the original data sets are damaged or accidentally deleted. On the days and at the time specified by the storage administrator, DFSMShsm automatically copies new and changed data sets on DFSMShsm-managed volumes to DASD or tape backup volumes.

The DFSMShsm copy is a backup version that resides in an optionally compacted, device independent format on a lower cost-per-byte storage device. Users can recover their own data sets without intervention by a storage administrator.

During the backup process, DFSMShsm prevents anyone from updating the data while the data set is being copied, unless the data set is backup-while-open eligible. If the data set is backup-while-open eligible, DFSMShsm invokes DFSMSdss to perform the incremental backup, but the data set change indicator is not reset, so the data set gets backed up again the next time backup is performed. DFSMShsm records the location of the backup version in its backup control data set and maintains records of its tape volumes in the offline control data set.

DFSMShsm can back up a maximum of 15 user volumes concurrently. Each incremental backup task selects its own daily backup volume, either tape or DASD. The storage administrator selects the maximum number of concurrent tasks.

DFSMShsm allows automatic retention of up to 100 backup versions of each data set. DFSMShsm has the capability to determine and delete those backup versions that are no longer needed according to criteria supplied by the storage administrator.

To protect your inventory of backed up and migrated data, the three DFSMShsm control data sets and journal are backed up as a synchronized set at the start of automatic incremental backup. DFSMShsm maintains a user specified number of backup and journal generations. The journal, a record of all transactions posted to the control data sets between backups, is used in the recovery process to update the recovered control data sets to the point of failure.

Figure 1 shows the backup paths of data sets managed with SMS and without SMS.

Figure 1. Automatic incremental backup

For system-managed volumes, DFSMShsm uses the storage group automatic backup attribute to determine what volumes to back up, and the management class attributes to determine whether to back up a data set and, if so, how often to back it up. After the data set is backed up, DFSMShsm determines from the management class attributes how many versions to maintain and how long to keep them.

For non-system-managed volumes, DFSMShsm uses the volume attributes to determine what volumes to back up. If DFSMShsm’s automatic backup attribute is active for the volume, all data sets requiring backup are backed up.